Many electrical, electronic, optical, magnetic, electromagnetic and electro-optical devices require single crystal-like device layers with few defects within the device layer for proper operation. For such applications, single-crystal or single-crystal-like devices offer the best performance. Fabrication of large-scale single-crystal-like devices can be accomplished by epitaxial growth of these devices on lattice-matched, single crystal substrates of ceramic materials of oxides, nitrides and carbides, such as MgO, Al2O3, SrTiO3, LaAlO3, NdGaO3, LaAlO3, YAlO3, and LSAT ((La0.18Sr0.82)(Al0.59Ta0.41)O3), Si, GaN, and SiC. However, these substrates cannot be fabricated using known processes in long lengths or in large areas and are generally limited to sizes of no more than about a foot in length and diameter. These substrates are also rigid and not flexible because these substrates are typically at least about 1 mm thick.
A variety of artificially fabricated, polycrystalline, single crystal-like (sometimes referred to as being sharply textured) substrates have been developed. Among them, an important class of substrates is known as rolling assisted, biaxially textured substrates (RABiTS). Biaxial texture in a substrate refers to a situation where the are a plurality of grains (thus being polycrystalline substrate), where the individual grains in the polycrystalline substrate are preferentially aligned within a certain angular range with respect to one another in two orthogonal directions. A polycrystalline material having biaxial texture of sufficient quality for electromagnetic applications, such as superconducting applications, can be generally defined as being characterized by an x-ray diffraction phi scan peak of no more than 20° full-width-half-maximum (FWHM) and a omega-scan of 10° FWHM. The X-ray phi-scan and omega-scan measure the degree of in-plane and out-of-plane texture, respectively. An example of biaxial texture is the cube texture with orientation {100} <100>, wherein the (100) crystallographic plane of all grains is parallel to the substrate surface and the [100] crystallographic direction is aligned along the substrate length.
A variety of buffer layers are known, generally for use with polycrystalline, biaxially textured substrates. The disclosure of the following U.S. patents is incorporated herein by reference for each of its disclosure regarding textured buffer layers: U.S. Pat. Nos. 5,739,086; 5,741,377; 5,846,912; 5,898,020; 5,964,966; 5,958,599; 5,968,877; 6,077,344; 6,106,615; 6,114,287; 6,150,034; 6,156,376; 6,151,610; 6,159,610; 6,180,570; 6,235,402; 6,261,704; 6,270,908; 6,331,199; 6,375,768, 6,399,154; 6,451,450; 6,447,714; 6,440,211; 6,468,591, 6,486,100; 6,599,346; 6,602,313, 6,607,313; 6,607,838; 6,607,839; 6,610,413; 6,610,414; 6,635,097; 6,645,313; 6,537,689, 6,663,976; 6,670,308; 6,675,229; 6,716,795; 6,740,421; 6,764,770; 6,784,139; 6,790,253; 6,797,030; 6,846,344; 6,782, 988; 6,890,369; 6,902,600; 7,087,113. Moreover, there are other known routes to fabrication of polycrystalline, biaxially textured, substrates, such as ion-beam-assisted deposition (IBAD) and inclined-substrate deposition (ISD). IBAD processes are described in U.S. Pat. Nos. 6,632,539, 6,214,772, 5,650,378, 5,872,080, 5,432,151, 6,361,598, 5,872,080, 6,756,139, 6,884,527, 6,899,928, 6,921,741; ISD processes are described in U.S. Pat. Nos. 6,190,752 and 6,265,353; all these patents are incorporated herein by reference for their IBAD or ISD related disclosure. In the IBAD and ISD processes a flexible, polycrystalline, untextured substrate is used and then a biaxially textured layer is deposited on this substrate. Large-area and flexible single crystal metal or alloy substrates have also been fabricated as reported in U.S. Pat. No. 7,087,113 by Goyal. U.S. Pat. No. 7,087,113 by Goyal is also incorporated herein by reference for its disclosure regarding large-area and flexible single crystal metal or alloy substrates.
Superconducting tapes based on epitaxial growth of superconductors on biaxially textured substrates described above using RABiTS, ISD or IBAD processes or based on the single-crystal metal or alloy substrates disclosed in U.S. Pat. No. 7,087,113 by Goyal are currently slated to be used for most large-scale applications of high temperature superconductors. The conductors based on these techniques use a metallic substrate in the form of a tape, typically a cm or so wide from which about 0.4 cm wide wires are slit. However, for all these conductors, the AC losses are generally high. AC losses result in the YBCO coated conductor as a result of either applying ac currents to the conductor or placing the conductor in applied alternating magnetic fields. Minimizing these losses entails reduction of hysteretic losses in the superconductor, the substrate normal metal effects such as eddy currents, ferromagnetic substrate contributions, and coupling current losses. Depending on the application environment, certain types of these losses dominate. A principal concern is generally the hysteretic loss, but the other losses can become quite important, especially at high frequencies. These losses are high because it is not possible to fabricate narrow wires. Hence, the successful incorporation of superconducting tapes into these ac-applications will require the development of an ac-tolerant version of the superconducting wire that will sufficiently minimize these effects.
Methods to reduce the ac losses are being developed. Ferromagnetic loss of the substrate can essentially be removed by the use of non-magnetic substrates, as is one of the goals for the RABiTS process. Reduction in hysteretic losses in the superconductor layer can be done by making the coated conductor with a filamentary design and making the filament width very narrow, such as about 100 microns or so. Eddy currents and coupling currents can be reduced by increasing the resistance of the substrate and the interfilamentary path, respectively. Twisting (or some scheme to allow field penetration), whether in the YBCO coated conductor architecture or of the conductor itself is generally necessary to further reduce AC losses. However, these techniques to reduce AC losses are quite difficult to implement with the epitaxial superconductor on RABiTS, ISD and IBAD or single-crystal metal or alloy substrates. Addressing the issues with the metal or alloy substrate is perhaps the simplest to do and is generally the only primarily an issue with the RABiTS approach.
However, reducing the hysteretic losses in the superconductor layer is far more difficult. First of all, filamentization of the superconductor layer in the tape conductor into very narrow filaments is necessary. This is not only quite difficult and cumbersome to do in kilometer long superconducting tapes but also adds a very significant cost to the fabrication process making the total cost of the conductor too expensive for most large-scale applications. Second, as noted above, twisting and transposing of the filaments in a tape conductor is necessary to further reduce AC losses. However, twisting of a tape conductor is quite difficult and may not be possible in a practical process. So far twisting has not been demonstrated convincingly for coated conductors. Hence, it is unclear if a superconducting wire based on epitaxial superconductor on RABiTS, ISD and IBAD tape or single-crystal metal or alloy substrates will find use in large-scale applications of superconductors wherein low AC-losses are required. These applications include underground transmission lines, motors, generators, and high-field magnets, etc.
Ideally, a “round” superconducting wire is desired for most large-scale applications of high temperature superconductors. Of course, this wire needs to be single crystal or single-crystal-like in its crystallographic orientation since high-angle grain boundaries suppresses or disrupts supercurrents. Ideally, this wire will not be a metal or alloy based wire so that both ferromagnetic losses and eddy current losses are not an issue. This “round” wire also needs to be flexible in that it can be bended and twisted to transpose superconducting wires together to minimize AC losses. While this has been a need in the field of superconductivity for about two decades, no method or route exists to fabricate such a “round” superconducting wire.
As noted above, in addition to the high-temperature superconductor applications, for many other applications, including, but not limited to, electronic, optical, magnetic, electromagnetic and electro-optical applications, single crystal-like device layers with few defects within the device layer are required. For many of these applications, metal or alloy substrates are not desired. Furthermore, single-crystal-like substrates do not suffice for many semiconductor device applications as the use of these results in a high defect density in the device layer. In addition, for many of these applications, thick, rigid, ceramic single crystals are presently being used as substrates for the devices.
For example, single-crystal sapphire or Al2O3 substrates are commonly used in solid state lighting applications. The single-crystal ceramic substrates used in these applications are thick, inflexible, and non-faceted, and made using standard crystal growth processes which are generally both slow and expensive. For many of these device applications, reduction in cost of the substrate as well as reduction in the cost of fabricating or depositing the device layers on the substrates is a major concern. Finally, for many of these applications, having a flexible substrate offers great potential advantages. For example a single crystal wire or tape-like solid-state lighting device would be ideal for lighting applications. Many sensors which operate in harsh environments could use a long single crystal or single-crystal-like substrate which contains the device at one end. In addition, for some applications it may be desirable to couple semiconducting properties with optical properties for short fiber lengths. For other semiconductor-based applications such as photovoltaics or solar cells, it may be desired to have a tape geometry wherein the tape is flexible.
However, there is no known relatively rapid and low cost method for fabricating single-crystal wires, fibers or tapes of these device layers. Moreover, the crystals provided by available single-crystal, structural, ceramic fibers and tapes or ribbons, such as sapphire, are generally highly inflexible based on their thickness, and are also incompatible with the crystal orientation required by commercially valuable device layers, such as superconducting layers. Therefore, what is needed is a new low cost process that permits rapidly forming thin single-crystal, ceramic fibers and tapes or ribbons that include features that permit epitaxial device layers to be grown thereon to support a variety of new improved epitaxial devices. Thin single-crystal ceramic fibers and tapes or ribbons would provide flexibility to allow twisting, braiding or transposing relative to one another along a length of the article, such as for reducing AC losses.